UX10: Capstone: three design projects for SehatSlot
Project 1: a study reported honestly
🎯You'll be able to
Complete the research project: plan, findings, counts and limits that pass all eight research tests.
🏠The everyday picture
A good court report states what each witness said, how many said it, and what the evidence can't prove. A judge can then decide fairly.
📘What it really is
-
Plan: rewrite the starter
research/study-plan.mdwith Goal, Participants (IDs only), Tasks (goals, not buttons), Consent (recording, stopping, deleting) and Measures. -
Findings: write
research/findings.mdin the format inresearch/README.md: Summary, Task results, Survey, Findings and Limits. - Evidence: each finding's tag must match the session notes exactly, with every participant who showed it and nobody else.
- Counts: completed without help, with help and failed, as
x/5; responses and the mobile share from the survey. - Changes: every critical and serious finding has a design change, which Projects 2 and 3 carry out.
📝A worked example
Input: the eight tags and the result lines.
Expected output (the reference findings):
| Tag | Evidence | Severity |
|---|---|---|
| slot-picker | P1, P2, P4, P5 | critical |
| error-colour | P1, P3, P5 | critical |
| no-back | P1, P3, P5 | serious |
| no-results | P2, P4 | serious |
| placeholder-label | P2, P3 | serious |
| small-targets | P3, P5 | serious |
| phone-shown | P4 | serious |
| jargon | P1 | minor |
Task results: T1 2/5, 2/5, 1/5; T2 3/5, 0/5, 2/5; T3 2/5, 1/5, 2/5. Survey: "Responses: 120" and "On mobile: 79" (a whole percentage: 95 of 120). Then npm run test:research
shows 8 passed.
🌍Where it's used
- Research reports that teams use to plan the next quarter.
- Accessibility audits that must trace each issue to evidence.
- Portfolios, where honest numbers show maturity.
📝Case scenario
A learner's findings said "most users struggled with errors". Test 5 failed: the notes showed [error-colour] for exactly P1, P3 and P5.
Rewriting the evidence to match the notes, and the summary to "3 of 5", made the report both passing and more convincing.
🖱️Try it
Run npm run test:research on your draft and fix one failure at a time, reading the test's message each time. Success: 8 passed, and no
test was changed.
⚠️Common mistakes
- Evidence that doesn't match. Sign: test 5 fails. Fix: list exactly the participants whose notes carry the tag.
- Merged counts. Sign: test 6 fails on "with help". Fix: three separate counts.
- Leading tasks. Sign: test 2 fails on "filter" or "tap". Fix: describe the goal.
🛠️For your project
The project page's marking guide adds two safety items: no personal details in research files, and consent covering recording, stopping and deletion.
✅Quick check
What must a finding's Evidence column list?
Show the answer
Every participant whose notes carry its tag, and only those.
How are task results written?
Show the answer
As three separate counts out of five: completed without help, with help and failed.